Recognizing individual social partners requires that they have variable, individually distinctive phenotypes. Despite the importance of recognition in mediating social interactions, relatively few studies have examined how the extreme phenotypic variation used for recognition evolves. Theoretical work suggests that variation in identity signaling traits is maintained by negative frequency-dependent selection, though no studies have directly examined the underlying genetics of identity signals to test this hypothesis. Here we propose to use the house mouse, Mus musculus domesticus, the premier mammalian model organism, to study the genetic basis of extreme variation in identity signaling phenotypes. Scent-based individual recognition in mice is mediated by distinctive major urinary protein (MUP) profiles in urine. MUPs are the products of a family of tandemly duplicated Mup genes on chromosome 4. Using the F1 male progeny of wild-caught mice, our study will examine the variation in MUP profiles in three ways. (1). Our first objective is to document patterns of DNA sequence variation at Mup loci in natural populations. Comparisons of inbred laboratory strains and our preliminary analyses suggest that distinctive MUP profiles result from a combination of allelic and gene copy number polymorphism. We will use patterns of DNA sequence variation to test for evidence of negative frequency-dependent selection maintaining MUP diversity. (2). We will examine patterns of Mup gene expression in wild-derived mouse progeny. Previous studies have indicated that Mup loci are variably expressed across inbred laboratory strains. We will document patterns of Mup expression within and between individuals. Different patterns of expression for the same Mup gene across individuals would suggest that lineage-specific differences in Mup regulation may be an important contributor to the highly variable phenotypes found in natural populations. (3). The relative importance of allelic, gene copy number and regulatory variation in producing distinctive phenotypes is unknown. We will determine the relative contributions of each form of genetic variation to distinctiveness by comparing urine profiles with genetic data for the same individuals. Additionally, we will test if protein variatio is recognizable as individually distinct by females, linking genes to proteins to behavioral phenotypes. This work will be done in collaboration with Drs Robert Beynon and Jane Hurst (University of Liverpool, UK), two recognized experts in these areas respectively.